Volatility Dispersion Presentation for the CBOE Risk Management Conference Izzy Nelken 3943 Bordeaux Drive Northbrook, IL 60062 (847) 562-0600 www.supercc.com www.optionsprofessor.com Izzy@supercc.com
Volatility Dispersion This is a term that s s been bandied about Typically means: I I have volatility exposure in one or more instrument, can I get rid of it using other instruments
Volatility tracking Market makers used to sell options on the US dollar / Swiss franc and hedge using the US dollar / German mark As long as both currencies moved in tandem, this worked rather well But there were periods where the currencies diverged and the hedge fell apart
Past correlation In the pervious example, there was a relationship between the currencies that has been observed in the past It does not mean that it will be sustainable in the future
Types of Equity Indexes Focus of Index Broad Index (S&P 500, Nasdaq) Narrow (industry group) index (e.g. Semiconductor index, Health Care index) Weighting Market Cap weighted (e.g. S&P 500, Nasdaq) Equally weighted price (Dow Jones) Equally weighted return index (e.g. the Value- Line index)
Index to component arbitrage Many try to arbitrage between the index (e.g. DJIA) and its components stocks Others also include futures (DJIA index futures) This relationship has been arb-ed to death
Relationship scale Simple relationship (e.g. index to components easy to do so there is not much profit potential) Historical relationship (very dangerous - high possible profits but also high potential risks) In the middle: relationship between the implied volatility of the index and the implied volatility of its components
Index options Some index options are only traded at a single exchange as compared with equity options which may be traded at up to five exchanges Index options have high bid-ask spreads as compared to single stock options Have high implied volatilities as compared with the historical volatility of the index
Implied vs. Historical NDX Avg Hist 41.43%, Imp 44.28% NDX : Imp & Hist Vols 80% 70% 60% 50% 40% 30% 20% Dec-01 Feb-02 Apr-02 May-02 Jul-02 Sep-02 Oct-02 Dec-02 Jan-03 Mar-03 May-03 Implied Vol Historical Vol
Implied vs. Historical GE Avg Hist 40.03%, Imp 40.13% 100% 80% GE : Imp & Hist Vols 60% 40% Implied Vol Historical Vol 20% 0% Dec-01 Feb-02 Apr-02 May-02 Jul-02 Sep-02 Oct-02 Dec-02 Jan-03 Mar-03 May-03
Hedge Fund Strategy - I Some hedge funds just sell options on the index: Sell at the money near term (1-2 2 months) S&P options Buy far from the money 1-21 2 months S&P options As the hedge fund says Natural suppliers of index options do not exist Buyers must pay a premium Benefit from overpricing on index options
Rationale Because the implied is higher than the actual volatility So they collect more premium than they risk to their capital Example: Acorn Derivatives
Hedge Fund Strategy - II Sell index options close to ATM straddles or strangles Buy component options
Why not the reverse? Short index options Long component options not the reverse Index options are expensive You can be surprised with a single stock (e.g. bankruptcy, takeover, earnings surprise) but not with the entire index
Delta hedging Each are Delta hedged with its own underlying As this is a hedged position They get a better treatment from the clearing firm It is really a volatility play or more precisely a correlation play
Examples The Barracuda fund of Artradis in Singapore Balanced Asset Management of Hungary Our fund (Keren( Asset Management) in the USA may also employ this strategy The Buy-Write index takes advantage of the same phenomena
Market Makers Are attracted by the high bid ask spreads in index options They sell these index options and benefit from high implied volatilities as well as the large bid-ask spreads Market makers may find it difficult to buy them (as natural sellers may be hard to come by) So they need to hedge their volatility exposure (vega)
Reducing vega By purchasing component options which are relatively cheap and have low bid ask spreads
High Volatility Typically, volatility rises when markets tumble Stock drop across the board Correlation is high Sell straddles in high volatility environments but worry: can volatility increase even more? Volatility dispersion is a way to reduce the risk Since correlation is bounded <=1
The Portfolio Equation n n n Var = 2 σ 2 w + ρ σσw w p i i ij, i j i j i= 1 i= 1 j= 1 i j σ = p Var p
Portfolio Equation - II σ = standard deviations of returns ρ = correlation of returns wi = weights for each security
Currency Options The portfolio equation is used by foreign currency (forex( forex) ) option participants Forex option triangle: Participants know the volatility on: USD/Euro options USD/JPY options Euro/Joy options From here, we can get at the correlation numbers
Two stock index Given the two implied volatility numbers and the index implied volatility We can deduce the correlation number this is the implied correlation If it is much higher than the historical correlation number it may be a good trade
Hypothetical Example Weight in Index Implied Volatility Stock A 70% 30% Stock B 30% 40% Correlation 70% Volatility of index 30.62%
Question Sold 100 index ATM straddles How many straddles in each component? Sold 100 index slightly out-of of-the-money call options How many call options in each component? Which strikes should you use?
Equating Greeks Delta s s are hedged in each instrument separately If the trader sold 5000 Gamma (or Vega) units on the index option they would like to have an equivalent amount in the component options
Profit If the implied correlation comes back in line with normal levels by index implied volatility dropping by component implied volatility increasing Then the trade makes money Windfall win when one stock jumps up and the other down The index is unchanged The component options made money
Danger As the stocks & index move these options may no longer be close to ATM So even if the implied volatility numbers came in line you may not have made any money Do you readjust the option position?
Correlation explosion The trade is a bet that correlation will return to normal levels If correlation continues to increase (e.g. due to all stocks moving in the same direction) the trade may, in fact, lose money
Broad vs. Narrow indexes Broad index has many constituents so must buy many component options and manage them Narrow index has few component options so it is an easier trade Sector index has high levels of correlation to start out with when it is really high there is not too much correlation risk (the most it can go to is 1)
Index weights As the stocks move up and down, their relative weightings in the index change This impacts on the index implied volatility If a stock moves down dramatically, then its weight in the index drops. On the other hand, its implied volatility increases These effects tend to partially cancel each other
Many stocks Weight in Index Implied Volatility A 10% 30% B 15% 40% C 20% 45% D 25% 40% E 30% 45% Correlation matrix A B C D E A 100% 70% 30% 60% 70% B 70% 100% 60% 50% 40% C 30% 60% 100% 60% 50% D 60% 50% 60% 100% 70% E 70% 40% 50% 70% 100% Index Volatility 33.918%
Difficulty With many stocks, we can not uniquely determine an implied correlation matrix One method Use the historical correlation and the implied volatility Compare the resulting volatility with the actual implied volatility of the index
How to compare? If the implied volatility of the index is 35% but the result of the computation is 33% - is that enough to be a seller? Need some way to assess the relative implied volatility differential (e.g. on a historical basis)
More difficulties We do not hedge with the entire index But with a select group of stocks These are the heavily weighted stocks large market cap the ones with large liquidity the ones with large implied volatility numbers How many stocks should you choose?
Lag Risk It is often impossible to place all the legs of this trade simultaneously.. so the trade is exposed to lag risk
Volatility dispersion models Ivolatility.com Brisk systems and others
Profit attribution Why did we make money on a particular trade? Was it a result of implied correlation coming in? Or perhaps the trader accumulating Deltas?
Science & Art We have developed volatility dispersion tools for consulting clients These models are quite mathematical But volatility dispersion trading is also an art Beware of a black box model